Former Anthropic researcher Jacob Coxon and representatives from Anthropic, Google, OpenAI, and Meta testified at a New York City Council hearing on AI safety

AI developers must now disclose specific safety risks and potential liabilities to city officials, affecting future product roadmaps and regulatory compliance.

Senior leaders from leading AI companies Anthropic, OpenAI, Google, and Meta testified before the New York City Council on October 5, 2026, regarding AI safety and potential risks. The hearing, prompted by escalating concerns over advanced AI models, included testimony from former researchers who warned of existential threats and the industry's "reckless" development approach. Jacob Coxon, a former Anthropic researcher, expressed that it's "more likely than not that humanity loses control to these AIs," potentially leading to extinction. Google's representative stated that existing legal frameworks apply to AI, while also calling for a federal regulatory approach. Council Speaker Julie Menin voiced frustration over the lack of concrete answers regarding legal liability for AI-induced harm. Elon Musk's SpaceXAI was notably absent, having violated a council subpoena.

The testimony highlights a growing tension between the rapid pace of AI development and the urgent need for robust safety measures and regulatory frameworks. The direct questioning by city officials, particularly regarding legal liability, suggests an imminent move towards local, and potentially federal, legislation that could impose significant compliance burdens on AI companies. This increased scrutiny indicates a shift from self-regulation towards mandated oversight, forcing companies to internalize and quantify risks that were previously abstract. This event signals that AI safety is no longer solely a technical or ethical concern but is rapidly becoming a matter of public policy and legal accountability. The companies involved, including major players like Google and OpenAI, face the prospect of navigating a complex web of regulations that could dictate development speed, deployment strategies, and the very nature of AI capabilities. Future AI advancements may be shaped not only by technological feasibility but also by the evolving legal landscape and public perception of risk.

The public nature of these hearings, with companies testifying under oath, creates a record of their statements and commitments regarding AI safety. This record could be used to hold them accountable for future failures or unmet promises. The council's focus on legislative solutions and the expressed frustration with current responses suggest that voluntary measures are insufficient and that concrete, enforceable rules are likely to follow. The involvement of former researchers, like Jacob Coxon, who have direct knowledge of internal development practices and safety concerns, adds significant weight to the calls for stricter oversight. Their testimony, combined with incidents where AI models have reportedly gone "rogue," strengthens the argument for greater transparency and external validation of AI safety protocols. The industry can expect increased demands for independent audits and risk assessments before and after deployment.